Gore1FL
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 02:38 PM
Original message |
|
Assuming Obama wins The Kerry and Gore States (which looks more than plausible as he leads in these states by over 10%)
He will only need one of the following battleground states to get 270+:
FL (Obama +3) Obama Win% 76% OH (Obama +6) Obama Win% 86% GA (McCain +4) Obama Win% 9% NC (Obama +2) Obama Win% 65% VA (Obama +7) Obama Win% 96% IN (McCain +2) Obama Win% 44% MO (McCain +1) Obama Win% 52% AZ (McCain +4) Obama Win% 4% CO (Obama +7) Obama Win% 96%
based on the combined winning%, Obama has a 99.9995581557637% chance of winning one of these states.
Or Any two of: MT (McCain +4) Obama Win% 12% ND (McCain +1) Obama Win% 19% NV (Obama +7) Obama Win% 84% SD (McCain +9) Obama Win% 1% WV (McCain +8) Obama Win% 1%
based on the combined winning%, Obama has a 25.796% chance of winning two of these states.
This brings us to a total of 99.9996721356703% chance of Obama winning 270+ Electoral votes through the combination of states above given the primary assumption (which gives him 264 Electoral votes to start with).
(margins from Nov 1 www.electoral-vote.com chances of winning each listed state from Oct 31 www.fivethirtyeight.com )
|
Celebrandil
(254 posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 02:46 PM
Response to Original message |
1. There must be a flaw in those numbers. |
|
Edited on Sat Nov-01-08 02:51 PM by Celebrandil
I suppose you are assuming the results from different states to be statistically uncorrelated. However, most likely they are in fact highly correlated. Thus the combined probabilities will be much different.
|
Gore1FL
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 03:07 PM
Response to Reply #1 |
2. I used standard methodology |
|
For example:
4% chance of winning AZ In the 96% chance that he loses it, there is a 9% chance of him winning GA (9% * 96%) = 8.64% 4% + 8.64% = 12.64% In the 87.36% chance that he loses AZ and GA, there is a 44% chance of him winning IN (44% * 87.36%) = 38.4384 + 12.64 = 51.0784%.
And so on. Remember he only has to win one of the first list of states. Hence each state % is based onthe percentage of him losing the state(s) listed above)
The methodology is right out of a Graduate level Statistics text book.
For example.
The probability of rolling a six on a single Die is 1/6. The chance of rolling a six on two single sided dice is 1/6 * 1/6. HOWEVER, if I need to roll a single 6 and have a pair of dice to do it with the formula is 1/6 + (1-1/6)* 1/6.
|
Celebrandil
(254 posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 04:57 PM
Response to Reply #2 |
4. Well... it's hard to model dependencies |
|
You did what people normally do and there is nothing wrong with that. If you can't model the dependencies, simply assume there are none.
However, it's easy to understand that if Obama wins Arizona, the probability of him winning Georgia is probably higher than 9%. Something fundamental has happened in Arizona, that is most likely true also in Georgia.
|
Gore1FL
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 06:42 PM
Response to Reply #4 |
|
My main reason for the thread was really to give a checklist of states to watch.
The stats was just something I had been playing with and decided to add.
Pretty much, if any of the states I listed in the top column are called, it is safe to gather the part favors!!
|
Gore1FL
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Nov-01-08 03:19 PM
Response to Reply #1 |
|
Yes I used straight statistics and did not correlate events that could effect the percentages enmass.
I know of no statistical formula available to correlate the many unknown variables needed to do so.
Nonetheless, the methodology is an acceptable model or probability based on the given win percentages in the states included.
|
DU
AdBot (1000+ posts) |
Tue Apr 30th 2024, 03:04 PM
Response to Original message |